Summary of "Valor en Riesgo Condicional (CVaR). Definición y ejemplo en R"

Valor en Riesgo Condicional (CVaR) - Definition and Example in R

This video explains the financial risk measure Conditional Value at Risk (CVaR), its relationship to Value at Risk (VaR), and demonstrates how to calculate it using R with a practical example based on gold price data.

Key Concepts and Finance-Specific Content

Risk Measures Covered

Value at Risk (VaR):
- A traditional risk metric estimating the maximum expected loss over a given time horizon at a specified confidence level (e.g., 5%).
- VaR focuses on the left tail (negative returns) of the return distribution but ignores the magnitude of losses beyond the VaR threshold.
- Example: VaR at 5% means there’s a 5% chance losses exceed this amount.
Conditional Value at Risk (CVaR) or Expected Shortfall:
- An improvement over VaR that averages losses exceeding the VaR threshold, capturing tail risk more effectively.
- More coherent and restrictive than VaR, providing a better measure of extreme losses.
- CVaR is always greater than or equal to VaR in absolute terms (i.e., more conservative).

Why CVaR is Preferred Over VaR

VaR ignores the severity of losses beyond the threshold, potentially underestimating risk.
CVaR accounts for the “tail” or extreme loss region, providing a fuller picture of downside risk.
Standard deviation and variance consider total volatility (positive and negative), which may not always reflect risk properly, especially for asymmetric return distributions.

Assets and Data Used

Gold price data from Yahoo Finance, from 2015 onward.
The video uses logarithmic returns of gold prices to calculate VaR and CVaR.

Methodology / Step-by-Step Framework for Calculating VaR and CVaR in R

Data Preparation:
- Download asset price data (gold prices).
- Calculate logarithmic returns to obtain a return distribution.
Visualize Returns:
- Plot returns and histograms to understand the distribution and tail behavior.
Calculate VaR:
- Use the quantile function to find the 5th percentile (or 95th percentile depending on perspective) of returns.
- This percentile represents the VaR at the chosen confidence level (e.g., 5%).
Calculate CVaR:
- Identify all return observations less than or equal to the VaR threshold.
- Compute the average of these “tail” losses to get CVaR.
- CVaR captures the expected loss given that losses exceed the VaR.
Interpret Results:
- VaR example: Gold VaR at 5% was approximately -1.5% return (meaning a 5% chance of losing more than 1.5%).
- CVaR example: Gold CVaR was about -2.1%, reflecting more severe average losses in the tail.
Alternative Perspective:
- Calculations can be done using the 95th percentile and positive returns by flipping signs, confirming equivalence of approaches.
Visualization:
- Plot the return distribution with lines marking VaR and CVaR to illustrate the differences and tail risk visually.

Important Numbers and Timelines

Gold price data from 2015 to present used for calculations.
VaR at 5% confidence level ≈ -1.5% return (loss).
CVaR at 5% confidence level ≈ -2.1% return (more conservative risk estimate).

Tools and Packages Mentioned

R language for data analysis and visualization.
Mention of packages like PerformanceAnalytics that have built-in functions for VaR and CVaR calculations.
The presenter opts to calculate manually to show the underlying methodology.

Recommendations and Cautions

CVaR is a more reliable and restrictive risk measure than VaR, especially for portfolios or assets with fat-tailed or skewed return distributions.
Understanding the difference between VaR and CVaR is crucial for proper risk management.
Calculations depend on the confidence level chosen (commonly 5% or 1%).
Clarification on percentile use (5% vs. 95%) is important to avoid confusion.
This content is educational and not financial advice.

Disclosures

The presenter clarifies the content is not personal invention but based on established research papers (links provided in video description). The video is for educational purposes, demonstrating statistical risk measures and R programming examples.

Presenters / Sources

The video is presented by an educator knowledgeable in financial risk measures and R programming (name not explicitly mentioned in subtitles).
References to Álvaro Alcívar for conceptual explanations.
Mention of external resources such as scientific papers and previous videos on VaR.

Overall, this video provides a clear theoretical and practical introduction to CVaR, highlighting its advantages over VaR, and offers a hands-on R example using gold price data to calculate and visualize these risk metrics.